re: #425 Charles

Exactly. This foolishness is almost beyond belief.

The problem I have with this sort of bilge is that there’s simply no way at all to deduce what an entire program does from a small fragment of code. In the case above, it’s even worse - the author is attempting to draw a conclusion from a single data element. We don’t even know the purpose of that element, let alone what role it plays in the subsequent routine, nor what role that routine plays in whatever program it is a part of.

It’s nearly identical to counting how many times the words “I” and “me” occur in a speech and deriving a psychological diagnosis of the speaker from that alone.

This particular author is extremely disingenuous in a number of other ways that fairly scream fraud and bias themselves. Apparently, it isn’t enough for him to simply force an interpretation that fits his needs; he seems to realize how weak his argument is, and performs all sorts of distortions, omissions and outright misrepresentations in order to shore up his position.

The way to check the code is: get all of it and run the original data through it to see if you get the same results as those published. Recreate the data from the raw samples and see it it matches. Analyze the filtering and scaling performed on the raw data to make sure it is sensible.

But don’t start making shit up that you know nothing about. It’s just too damn easy to detect, and it weakens your case by making you appear desperate.